sq. in. of surface area. While this is barely adequate for a well cooled computer it would not be adequate for most video games applications. Adtech Power uses an area adequate to provide 0.32 Watts/sq. in. and a maximum rise of 40°C. at 120 volt input. This is adequate to provide at least a 15 % margin on our transistor junction temperatures even at a 40°C. outside ambient and up to a 40°C. internal ambient temperature rise.
To design our theoretical power supply, we, of course, must know the required output voltage and current, the required line voltage range, and the maximum ambient temperature in the vicinity of the power supply. And of course, the regulation required.
For instance, T2 L chip's will tolerate a voltage change of ±5%, however, load wiring drop will generally be about 1–2%, on long wire system it could be close to 5%. The temperature drift of the better power supplies will be .02%/°C. and on the cruder discrete supplies as much as 0.1%/°C. which will cause another 1.1% to 5% deviation for a 40°C. ambient and 40° rise. Another 1–2% should be allowed for metering errors. PC Card and connector drops will account for 0.1–0.25% deviation. Ripple will cause a further deviation of 0.1–0.25%. So to maintain a ±5% reg. at the IC, our power supply usually has to maintain better than 1% overall line plus load regulation which is no major problem.
Our first requirement is to determine the required unregulated DC voltage. If an IC regulator is employed, the input voltage to the IC generally must be 3 volts above it's output voltage. Since on power supplies above 1 amp, a Darlington output stage will be required to keep the IC below safe dissipation area at high temperature, the IC output will have to be about 1.3 volts above the Darlington output voltage. Generally an additional 0.6 volts will be required for our current sensing resistor and 5% of the output voltage will be required to permit adjustment to compensate for output line drops, etc. In addition, the unregulated voltage has a considerable ripple content if we limit ourselves to economically practical filter capacitor size. This ripple will generally be 4–5% RMS or 5–7% peak. Since our regulator will drop out of regulation on the negative peaks of the ripple, our minimum voltage must be set above the required minimum DC input by that amount. Our power transformer will have an I2 R loss of 3–5% and since the resistivity rise of copper is 0.39%/°C., when the transformer is running at it's maximum temperature (about 100°C.), we must allow for this additional 29% of 5% drop or an additional 1.5%. Since all but the output voltage must be dissipated in the supply, it can be seen that for a 5 volt supply, the efficiency is extremely low especially when V in. min. is multiplied by 1.11 for nominal line. This would give us less than 34% efficiency when the rectifier losses and transformer losses are considered. To improve this situation, Adtech Power uses a boost supply to provide the input to the IC regulator only, permitting the main supply to go down to the saturation voltage of the power transistor. In this case, our dissipation is reduced by nearly 25%. The worst case dissipation at high line 110% load will be Pmax= [(V in. nom. x 1.1) - V out x]
–82–
